Understanding the local spatial and temporal patterns of attitudes towards immigration using Twitter data

Recent years have seen a growing interest in public attitudes towards immigration. Data derived from social media are increasingly being used to study attitudes beyond the limits of traditional data sources. In this paper, we utilise a dataset of 800,000 immigration-related tweet posted from Great Britain between 2013 and 2022. We analyse the sentiment of these, show temporal trends in sentiment and map areas with high relative levels of both strongly positive and strongly negative tweets. We then model the association of demographic and contextual variables with areas producing high proportions of strongly positive and strongly negative sentiment towards immigration.